An Interactive Kinect-Based Game Development for Shoulder Injury Rehabilitation

Abstract

Shoulder injury is a joint dysfunction that often occurs due to training or other activities that use hand as a support. Technology development such as motion tracking that can record movements becomes media that can be implemented in the rehabilitation process. An implementation that can be used to detect motion is by utilizing Microsoft Kinect. In this paper, we conducted research to utilize Kinect in an effort to assist the rehabilitation process, which can perform on the development of the players. We use several methods to predict and manage each level of the game, namely Exponential Smoothing and Kalman Filter. The results showed that the prediction with Kalman Filter was 87.59%, and the Exponential Smoothing method had an accuracy of 91.01%.